Projects

---

1. NASA Space Rock Watch

Developed a near real-time pipeline in Python orchestrated with Docker, Apache Airflow, Hadoop (HDFS, Hive), and PostgreSQL to fetch, store, and analyze NASA’s Near Earth Object API data. Implemented automated email alerts triggered when objects met specific criteria, enabling timely insights into asteroid/comet size, velocity, and close-approach trends for scientific research and public awareness.

⚙️ Tech Stack:

• Programming Language: Python
• Containerization: Docker
• Workflow Orchestration: Apache Airflow
• Big Data: Hadoop (HDFS, Hive)
• Relational Database: PostgreSQL

🖥️ Git - Source Code

2. Open Food Facts Data

Build a Python-based data pipeline to ingest and transform Open Food Facts API data using MongoDB, AWS S3, Snowflake, dbt and Metabase for BI dashboards. Orchestrated end-to-end workflows with GitHub Actions, providing actionable insights into product nutritional content and promoting better consumer health awareness.

⚙️ Tech Stack:

• Programming Language: Python
• API: Open Food Facts API
• Databases & Data Storage : MongoDB, AWS S3, Snowflake
• Data Visualization: Metabase
• Data Transformation & Modeling: dbt
• CI/CD & Orchestration: GitHub Actions

📊 Metabase - Dashboard Preview
🖥️ Git - Source Code

3. Know Me AI Bot

This interactive chatbot leverages OpenAI’s GPT-4o models and a straightforward file-based configuration to highlight personal details, skills, and experiences—no database required! It’s perfect for creating a personalized assistant or a dynamic portfolio showcase. Additionally, its 24/7 availability lets recruiters or anyone interested learn more about you anytime, saving valuable time on both sides.

✨ Highlights:

• GPT-4 Powered: Provides intelligent and context-aware responses.
• File-Based Configuration: Easy setup and customization without the complexity of database management.
• 24/7 Availability: Always online so recruiters or stakeholders can discover your profile on their schedule.
• Straightforward Deployment: Hosted on Render with minimal overhead.

⚙️ Tech Stack:

• Backend: Python (Flask)
• Frontend: HTML & CSS
• Model: OpenAI GPT-4o
• Hosting: Render

🤖 Chat with My Bot Assistant
🖥️ Git - Source Code

4. CoinGecko Cryptocurrency Market Data Pipeline

This project is a data pipeline that fetches, processes, and visualizes cryptocurrency market data from the CoinGecko API. It provides insights into trending cryptocurrencies, market trends, price movements, and more, leveraging data visualization using Tableau to aid in strategic decision-making.

✨ Highlights:

1. Enhanced Decision-Making: Get up-to-date market trends for data-driven investment and analytical strategies.
2. Time Efficiency: Automates data collection, eliminating the need for manual updates.
3. Comprehensive Insights: Aggregates multiple data points into a single pipeline for easier market analysis.
4. Improved Reporting: Provides real-time insights via Tableau dashboards.

⚙️ Tech Stack:

• Programming Language: Python
• API: CoinGecko API
• Database: BigQuery (SQL)
• Data Visualization: Tableau
  • Note: This project uses Tableau Public to keep it free. Tableau Public cannot connect directly to databases like BigQuery. Instead:
    • Google Sheets will be used to connect with Tableau Public for demonstration purposes.
    • BigQuery will be used to store and maintain the data, including historical updates.

📊 Tableau - Access the Dashboard
🖥️ Git - Source Code

5. Generalized Recommendation System

This project focused on building a comprehensive recommendation system pipeline—covering data preprocessing, model training, and performance evaluation. It can be adapted for recommending books, movies, products, and more.

✨ Highlights:

1. Scalable Recommendations: Flexible framework for various domains like e-commerce, entertainment, and education.
2. Enhanced User Experience: Personalized suggestions boosting user engagement and satisfaction.
3. Data-Driven Insights: Analyzes user behavior patterns to inform strategic business decisions.
4. Reusable Architecture: Modular design easily integrates with different datasets and use cases.

⚙️ Tech Stack:

• Programming Language: Python
• Libraries & Frameworks: Pandas, NumPy, Scikit-learn
• Preprocessing & Model Development: Jupyter Notebook
• Analysis Visualization: Matplotlib, Seaborn

🖥️ Git - Source Code

6. E-Commerce Data Workflow and Reporting Automation

This project is an automation framework designed to streamline data workflows and reporting processes in e-commerce. It facilitates data synchronization, processing, and report generation, ensuring efficiency and accuracy throughout the entire lifecycle.

✨ Highlights:

• Time Savings: Automates repetitive data processing and reporting tasks, minimizing manual effort.
• Improved Accuracy: Reduces human error in handling critical e-commerce data such as income, orders, and wallet payments.
• Enhanced Reporting: Delivers timely insights via automated report generation, enabling more effective data-driven decisions.

⚙️ Tech Stack:

• Programming Language: Python
• Storage: Google Drive (for saving the raw files to be processed)
• Database: BigQuery (SQL)
• Data Visualization: Google Looker Studio

📊 Looker Studio - Access the Dashboard
🖥️ Git - Source Code

7. E-commerce Store Metrics Dashboard

This interactive dashboard showcases consolidated data from multiple E-commerce platforms like Tokopedia, Shopee, and Lazada. It provides a unified performance view using Google Looker Studio. The dashboard is fully interactive, allowing users to filter by dimensions, adjust time periods (daily to yearly), and focus on specific metrics for actionable insights.

✨ Highlights:

• Business Overview: Track key store-level performance metrics such as revenue, visitors, conversion rates, average order value, and other key indicators essential for evaluating overall performance.
• Inventory Report: Monitor product-level metrics such as inventory value, aging, expired and damaged stock.
• Case Management: Evaluate customer service team performance in handling customer issues.

⚙️ Tech Stack:

• Data Visualization: Google Looker Studio

📊 Looker Studio - Access the Dashboard

8. Smart Sizing: Packaging Efficiency Enhancement

Imagine running a bustling online store, only to discover that oversized boxes and misalignment in packing can quickly erode your profits. This project tackles exactly that challenge—deploying Python scripts to verify the correct box usage for each order and identify potential alternative box sizes that fit more snugly. The result? Reduced overhead, less material waste, and a streamlined fulfillment process.

✨ Highlights:

• Automated Box Verification: Compares the warehouse’s chosen box size against each product’s actual dimensions, flagging oversizing and potential cost discrepancies.
• Cost & Sustainability Savings: Minimizes both excess shipping expense and environmental impact by cutting down on unnecessary packaging.
• Alternative Box Matching: Searches a list of different box dimensions (and costs) to find a more optimal fit—enabling future expansions with larger libraries of potential box sizes.

⚙️ Tech Stack:

• Programming Language: Python
• Libraries: Pandas, py3dbp (Packer, Bin, Item)
• Platform: Kaggle Notebooks

🖥️ Kaggle - Source Code

9. Leveraging Market Basket Analysis: Unearthing Hidden Gems in Sales Data

Dive into the world of Market Basket Analysis to unlock hidden purchasing patterns and enhance customers shopping experience. Utilizing algorithms like Apriori, we can identify which products are frequently bought together—providing insights for cross-selling, targeted promotions, and strategic product placements. Alongside this, exploring items purchased in bulk unveils further opportunities for volume discounts and seasonal deals.

✨ Highlights:

• Data-Driven Bundling: Spots products frequently co-purchased to guide effective cross-selling/upselling strategies.
• Strategic Promotions: Measures the strength of item relationships—via metrics like lift, confidence, and support—to tailor more compelling promotional offers.
• Bulk Purchase Insights: Analyzes multi-quantity purchases, opening possibilities for volume-based discounts, stocking strategies, and holiday sales.

⚙️ Tech Stack:

• Programming Languages: Python
• Libraries: Pandas, NumPy, mlxtend (Apriori & Association Rule Mining)
• Platform: Kaggle Notebooks

🖥️ Kaggle - Source Code

---